Fat Tails and Skewness into the Asset Allocation Decision

We investigate the impact of mean-conditional value-at-risk (M-CVaR) optimizations that take into account fat tails and skewness on optimal asset allocation. In a series of controlled optimizations, we compare optimal asset allocation weights obtained from the traditional mean-variance optimizations with those from M-CVaR. To arrive at efficient asset allocation weights that are different from the weights obtained from the traditional mean-variance optimization framework, the M-CVaR estimate must go beyond the first two moments, the mean and variance of the forecasted return distribution, and account for higher moments. We find that both skewness and kurtosis (fat tails) impact the M-CVaR optimization and lead to substantially different allocations than the traditional mean-variance optimizations. In particular, the combination of negative skewness coupled with a fat tail has the greatest impact on the optimal asset allocation weights. The study provides useful insights fo